Conversion Value Attribution Using User-Modified Baseline Curves
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional attribution models fail to account for future conversions related to an initial conversion, leading to an underestimation of the true value of impressions, as they only consider the most recent impression and not subsequent user actions.
Innovation Solution
An online concierge system predicts future user behavior by generating a baseline curve based on experimental data and modifying it with user-specific features to accurately estimate long-term incremental conversion values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional attribution models are used to link conversions to impressions, then the measurement process is simple and fast, but the measurement precision is insufficient because future conversions are not accounted for
Solution Approach 1:
The system performs preliminary actions by generating a baseline curve that predicts future conversion behavior before the actual conversions occur. This baseline curve is created using experimental data and user feature modifiers, allowing the system to anticipate and account for future conversions when measuring impression value, thereby improving measurement precision without requiring complex real-time analysis of every future event.
Solution Approach 2:
The system incorporates feedback mechanisms by using experimental data from other users to train the baseline curve and modifiers. This feedback loop allows the system to continuously improve its prediction accuracy by learning from observed user behaviors and adjusting the baseline curve accordingly, resolving the contradiction between simple measurement and precise measurement of impression value.
2Measurement precision
If only the most recent impression is attributed to the conversion, then the attribution process is simple and quick, but the true value of the initial conversion is underestimated
Solution Approach 1:
The system generates a baseline curve in advance that predicts the sequence of future conversions. This preliminary prediction allows the system to attribute conversion values to impressions more accurately without waiting for each future conversion to occur, thereby reducing the time loss associated with traditional attribution methods while improving measurement precision.
Solution Approach 2:
The system dynamically adjusts the attribution model by generating user-specific modifiers based on feature information. This dynamic adaptation allows the system to accurately capture the unique behavior patterns of each user, improving the attribution of conversion values to the appropriate impressions while managing the time required for accurate measurement through efficient computational methods.
3Measurement precision
If a baseline curve is generated for each user based on feature information, then the prediction accuracy improves, but the computational complexity increases
Solution Approach 1:
The system segments the prediction process into distinct components: a universal baseline curve generated from experimental data and user-specific modifiers derived from feature information. This segmentation allows the system to maintain high prediction accuracy by customizing the baseline for each user while managing computational complexity through modular processing of baseline generation and modifier calculation separately.
Solution Approach 2:
The system changes parameters by generating modifiers that adjust the baseline curve based on user feature information. These parameter changes allow the system to adapt the prediction model to individual users without requiring complete redesign of the entire prediction system, thereby improving prediction accuracy while controlling the increase in computational complexity through efficient parameter adjustment rather than structural overhaul.
Data Source
AI summary
An online concierge system generates the value for an impression by predicting future behavior by users beyond a current conversion. The predicted future behavior attributes incremental value of subsequent conversions by the user. The online concierge system gathers feature information about the user. Based on experimental data, the online concierge system generates a baseline curve describing expected user behavior for a category of users. Based on feature information of the user, the online concierge system applies a computer model to generate modifiers for the baseline curve to customize the baseline curve for the user. The modified curve is used to predict future actions by the user, and consequently a long-term incremental conversion value for the impression.